According to Kotler and Keller (2016), effective market research involves six steps: (i) definition of the problem and research objectives; (ii) development of the research plan; (iii) data collection; (iv) data analysis; (v) results' presentation; and (vi) decision making.
When researchers perform step (i) definition of problem and research objectives, that is, when the reasons that support the need for the development of field research and what is intended to obtain from the answers to the research are elucidated, it is possible to distinguish between three types of projects: exploratory, descriptive or causal (Hair, et al., 2008). Furthermore, during the definition of the problem and the objectives, the researcher defines the conception of the study, that is, whether it will be qualitative or quantitative (Malhotra, 2014).
For previously referenced authors, exploratory type surveys should be conducted when the objective involves gathering preliminary data to clarify the nature of the problem and suggest possible hypotheses or new ideas. Descriptive research should be used when describing market characteristics or functions, while causal studies are necessary when the objective is to test cause-effect relationships.
Having defined the type of research to be executed, it is possible to (ii) develop the research plan. This step establishes the approach and the sampling plan to be used, as well as the instrument and the most appropriate procedure for data collection (see item 2.1). After that, the (iii) data collection stage is characterized by requiring more time, as well as by the greater probability of errors, especially when the approach that is being used is the survey. In order to minimize the possibility of deviation in the results due to biased data collection, it is suggested to detail the collection procedure plan (Ray and Tabor, 2003).
Step (iv) data analysis involves the interpretation of the data collected. The methodology of data analysis to be performed depends on the initial objective of the research, as well as the type of project (Malhotra, 2014). In the presentation of the results (v), Kotler and Keller (2016) emphasize that it is the responsibility of the researcher to select the results that respond more specifically to the problem that led to the study, using appropriate communication techniques and describing conclusions relevant to the decision making. Regarding decision making (vi), when receiving the results of a market survey, managers should question whether the research carried out was based on an adequate scientific method (Clancy and Krieg, 2000) and should also reconcile the process of decision making with other data that has relevant business and environmental information, complementing the information generated in field research (Little, 1979).
2.1 Research Plan Development
According to Malhotra (2014), the research approach may be indirect or direct. In the indirect approach, the objectives of the project are masked for the respondents, while in the direct approach they are obvious to the participants considering the nature of the instrument. In other words, in the indirect approach techniques, the respondents provide a response that would not clarify the researcher if they were being directly questioned about a subject, whereas in direct approach techniques it becomes easier to obtain complete and detailed information. Regarding the means to be used, there are projective techniques and observation for the indirect approach (Anzieu and Chabert 2004; Underhill, 2008); and focus groups, individual interviews, surveys, and experimentation for the direct approach (Kuhfeld et al., 1994; Greenbaum, 1998; Bradburn et al., 2004; Giannarou and Zervas, 2014) as shown in Table 1.
The step called sampling plan requires three decisions to be made by the researcher: sampling unit, sample size, and sampling procedure. The sampling unit contemplates the definition of who will be researched, which is the target population that must be sampled. The decision on sample size signals how many subjects should be surveyed, and often samples with less than 1% of the population can provide results with a good level of credibility, once the sampling procedure chosen is reliable (Churchill and Iacobucci, 2006; Kotler and Keller, 2016). The sampling procedure indicates how the participants of the collection will be chosen. Representative samples should be taken from probabilistic samples of the population since in these samplings the calculation of confidence limits for the error is performed. However, when the monetary cost or time required to perform the collection is excessive, it is possible to use non-probabilistic samplings, although these do not allow measurement of the error (Hair, et al., 2008), as shown in Table 2.
Concerning the data collection instrument, it has a direct relationship with the design and the approach that will be used in the research. This means that in designing the data collection instrument, in addition to the main objective of the study, it should be considered whether it is qualitative or quantitative, as well as whether in-depth interviews focus groups, surveys, or other approaches (Churchill and Iacobucci, 2006; Hair et al., 2008; Malhotra, 2014), as summarized in Table 3.
Important care concerning data collection instruments is validation. The literature recommends that the instrument should be validated by specialists or that pre-tests are performed with a group of subjects with characteristics similar to those of the target population. Besides regard to structured questionnaires, it should also be considered the measurement and scheduling mechanisms that will be used (Bearden and Netemeyer, 1999), presented in Table 4.
Finally, the research plan should detail the methods of contact, that is, which will be the data collection procedures used with the elements of the sample. Given the type of approach, researchers may choose to collect data in person (projective techniques, focus groups, individual interviews, observation, survey, and experimentation) through postal methods (survey) or electronic methods (survey and observation). All methods have advantages and disadvantages depending on variables such as flexibility of the instrument, diversity of the questions, use of physical stimuli, subjects' anonymity, available time, and rate of return, among others (Churchill and Iacobucci, 2006; Malhotra, 2014).
2.2 Data Analysis
The form of data analysis that will be used stems from the research objective, the type of project that was carried out, and its design. Thus, qualitative exploratory research will have the results submitted to content analysis (Bardin, 2013), while quantitative descriptive and causal surveys will have the data obtained from the collection submitted to statistical analysis (Hair, Et al., 2010).
As qualitative research generates a significant amount of research notes or testimonials, content analysis suggests that the frequency of phenomena is counted, seeking to identify relationships among them, and the interpretation of the data is systematized according to conceptual models previously defined by the researcher. In addition to content analysis, qualitative research data can also be submitted to conversation analysis and semiotic analysis (Pierce, 1991; Goulding, 2005), as suggested by the information in Table 5.
On the other hand, the statistical analysis of the data coming from quantitative research usually goes through the following stages: univariate analysis, bivariate analysis, and multivariate analysis. In the univariate analysis, frequencies are established for each question researched. In the bivariate analysis, cross-tabulations are performed, and it is possible to calculate different measures of associations between the variables. Finally, a multivariate analysis should be applied when researchers wish to simultaneously analyze multiple measures on each subject or object being investigated (Hair, et al., 2010). Considering the strong use of multivariate analysis techniques in marketing research, Table 6 briefly seeks to describe them.